The present invention relates to a method and system for optical surveillance of a wide angle or panoramic field of view.
In certain imaging applications, an extremely wide field of view (120° or more) optical system is required so that a very large two dimensional region of the object space may be monitored either continuously or repeatedly at short intervals. Examples of such application include: full earth surveillance from low altitude space platforms, missile launch warning from an airborne platform, and air-borne threat detection from a ground base location or a waterborne vessel.
One approach to monitoring such large fields of view is the use of a scanning linear detector array. Examples of this approach are described in patent publications U.S. Pat. No. 5,347,391 and EP 1416312 A1. Although such systems offer a cost-efficient solution for scanning large regions, they suffer from a number of shortcomings. Most notably, a scanning linear detector array by its very nature actually views each given pixel of object space for a very small proportion of each scanning cycle. As a result, there is a significant risk of transient events, such as the brief flash accompanying the launch of a missile, being missed between scans of the sensor.
An alternative approach is to use staring imaging sensors to monitor the region of interest. Examples of systems employing staring imaging sensors include patent publications U.S. Pat. No. 6,410,897 B1, U.S. Pat. No. 5,534,697 A and U.S. Pat. No. 5,300,780 A. In most cases, in order to achieve acceptable resolution and avoid problems caused by optical distortion, the field of view of each imaging sensor should be limited to 40-60°. In order to cover a larger solid-angle field of view, a scanning pattern is typically used, resulting in similar problems as described in the context of linear detector arrays discussed above. For truly continuous non-scanned monitoring of a large field-of-view at an acceptable resolution, a large number of imaging sensors deployed with overlapping fields-of-view would be required, thereby rending the system very expensive.
One non-limiting example used to illustrate the present invention is that of protection of naval platforms (waterborne vessels or ships). Ships are relatively vulnerable to attack by many kinds of missiles, such as sea-skimming missiles and gliding bombs, and successful deployment of various countermeasures for their defense is dependent upon early detection of incoming threats. Radar has for decades been the standard technique for search and tracking of airborne threats for naval and other air defense systems. Radar is problematic, however, since it requires active transmission of radio pulses which give away the presence of the vessel carrying the tracking system and may be used as a guide beacon to guide offensive armaments towards the vessel. To avoid this problem, attempts have been made to develop passive (i.e., non-transmitting) search and tracking systems based upon optical sensors, and in particular, infrared search and tracking (IRST) systems.
Naval applications highlight the aforementioned shortcomings of both scanning and staring systems. Implementation of IRST systems for naval applications poses particular problems in the trade-off between sufficient sensitivity and avoidance of false alarms. Air-borne threat detection requires an extremely wide field of view, typically covering an azimuth of substantially 360°. During the short time-on-target (or “dwell time”) of a scanning system, the background optical noise of solar glint from the moving surface of the water is almost indistinguishable from the heat emission of a head-on incoming missile. As a result, scanning systems tend to suffer from insurmountable problems of high false alarm rates. If a staring system is used, although the dwell time problem is solved, a different problem of resolution vis-à-vis cost limitations arises. In order to achieve reliable detection of a head-on missile threat at sufficient range to be useful, an angular pixel resolution of at least two, and preferably at least four, pixels per mille-radian is required. To achieve this resolution with conventional imaging arrays of several hundred pixels dimensions, as many as 40-80 imaging sensors would be required, rendering the system overly expensive.
In other contexts, it has been proposed to use a single imaging sensor with optical multiplexing to perform more than one imaging function. Examples include the aforementioned U.S. Pat. No. 6,410,897 B1 where a movable mirror is used to switch the optical sensor between a wide field of view optical objective and a narrow field of view optical objective. A similar concept of switching between narrow and wide fields of view is also disclosed in U.S. Pat. No. 5,049,740, U.S. Pat. No. 4,486,662 and U.S. Pat. No. 3,804,976. Another example disclosed in U.S. Pat. No. 4,574,197 provides a scanning rotating polygon which offers two fields of view used for stereoscopic viewing or for two independently steerable optical telescopes for display on separate screens. None of these references discloses optical multiplexing to offer two similar fields of view with different optical axes in fixed spatial relation as a solution for a staring surveillance system.
A further limitation of the aforementioned existing systems with optical multiplexing is that the optical switching frequency is typically limited by the read cycle rate of the sensor, i.e., the period taken to expose the array to incoming illumination and then read the resulting information from an array of capacitors associated with each sensor element. In order to avoid mixing of the content of the two images, the sensor array is exposed for a first integration time to the first field-of-view, the associated capacitors are read (a first read cycle), and then the sensor array is exposed to the second field of view and the capacitors are again read (a second read cycle). This mode of operation is referred to as “Read Then Integrate (RTI). For surveillance applications in which it is desired to detect transient events of duration similar to or less than the read-cycle of the sensor, this arrangement is problematic since an event may occur while the other field of view is being viewed and may therefore be missed by the sensor.
Finally, in the field of staring sensors with a single field of view, there exists a technique known as “Read While Integrate” (RWI) which substantially avoids dead-time during the output reading process between integration periods of a sensor. This technique is particularly useful when monitoring for transient flash events since it helps to ensure that even a transient event is picked-up by the sensor. “Read While Integrate” also effectively doubles the rate at which frames can be acquired using an array of light-sensitive sensors that produce electrical charge when the light that they are sensitive to impinges on them. The principle of RWI will now be illustrated with reference to FIGS. 1 and 2.
Specifically, FIG. 1 shows one such sensor 10, for example an InSb detector sensitive to infrared light, coupled alternately to two capacitors 12 and 14 by a switch 16. Capacitors 12 and 14 in turn are alternately coupled to a readout circuit 18 by a switch 20. When a capacitor 12 or 14 is coupled to sensor 10, the capacitor 12 or 14 receives and accumulates (“integrates”) the electrical charge produced by sensor 10 as a consequence of the light impinging on sensor 10. When a capacitor 12 or 14 is coupled to readout circuit 18, readout circuit 18 reads the charge accumulated in the capacitor 12 or 14 and discharges the capacitor 12 or 14.
FIG. 2 shows the sequence of integration and readout used in the prior art RWI method to acquire images using an array of sensors 10 coupled to respective capacitors 12 and 14 and respective readout circuits 18 as illustrated in FIG. 1. Time increases from left to right in FIG. 2. During odd-numbered time intervals, capacitors 12 accumulate electrical charges while readout circuits 18 read the electrical charges accumulated in capacitors 14 during the immediately preceding time intervals. During even-numbered time intervals, capacitors 14 accumulate electrical charges while readout circuits 18 read the electrical charges accumulated in capacitors 12 during the immediately preceding time intervals. The read cycle, i.e., the period between successive readings from the same capacitor, corresponds to a pair of time intervals. The diagonal arrows in FIG. 2 show the timing of the flow of accumulated electrical charge from the capacitors 12 or 14 to readout circuits 18. Note that FIG. 1 illustrates the settings of switches 16 and 20 during odd-numbered time intervals.
Although RWI provides an effective solution for substantially continuous monitoring of an imaging system field of view, it is of limited value where a single imaging sensor is used to switch between two or more fields of view since each field of view would still remain unmonitored for at least half the read cycle.
There is therefore a need for a wide field-of-view surveillance system based upon staring imaging sensors which would employ optical switching to provide quasi-continuous monitoring of a wide-angle field of view while requiring fewer imaging sensors than would otherwise be required for full field of view coverage. It would also be advantageous to provide an infrared search and tracking system which would provide an effective passive alternative to radar for detecting threats to platforms, such as waterborne vessels.